Introduction: While patient throughput and emergency department (ED) length of stay (LOS) are recognized as important metrics in the delivery of efficient care, they must be balanced with the educational mission of academic centers. Prior studies examining the impact of learners on throughput and LOS when staffing directly with attending physicians have yielded mixed results. Herein we sought to examine the impact of a staffing model involving a supervisory resident "pre-attending" (PAT) on ED throughput and LOS, as this model offers a valuable educational experience for residents, but may do so at the expense of operational efficiency.
Methods: We retrospectively analyzed 26,702 unique patient encounters at a university-affiliated community ED between July 1, 2017-January 1,2019. The experimental group was comprised of patients seen primarily by midlevel providers, who staffed with a PAT, who subsequently staffed with an attending physician. The control group was comprised of patients seen by midlevel providers and staffed directly with attendings without a PAT. We used a parametric hazard model to analyze the effect of the presence of a PAT on service time, controlling for potential confounders including timing of presentation and patient demographics.
Results: The presence of a PAT is associated with a statistically significant increase in service time of five minutes (p = 0.006). Holding other variables equal, predicted service time in the experimental group was 173 minutes (95% confidence interval (CI), 171-176), while that for controls was 168 minutes (95% CI, 165-171).
Conclusion: The presence of a PAT is associated with a statistically significant increase in service time, but the magnitude (five minutes) is likely operationally insignificant. The negligible increase in service time is offset by the benefit to residents' training. The results of this study may be helpful for residency programs considering the addition of a PAT shift structure.